On the nonlinear output regulation for systems described by Takagi-Sugeno fuzzy descriptor models with a steady-state mapping as an LMI optimization problem

This paper is devoted to provide a numerical solution the nonlinear output regulation problem for descriptor systems. The control law under design is a nonlinear one, it consists on a nonlinear stabilizer combined with linear steady-state mapping as well as nonlinear steady-state input mapping; all of them are computed via linear matrix inequalities. A numerical example as well as a mechanical system as well are used to illustrate the viability of the proposed approach.

Fuzzy Descriptor Models for Earthquake Time Prediction Using Seismic Time Series

In recent years, singular systems and fuzzy descriptor have attracted a lot of researchers' attention due to their wide practical applications for modeling complex phenomena. In this study, one approach proposed is the fuzzy clustering algorithm based on linear structures to identify the neuro Fuzzy local linear models. Additionally fuzzy descriptor models, a recently proposed neuro fuzzy interpretation of locally linear models, are implemented because of their promise for intuitive incremental learning algorithms e.g. Generalized Fuzzy Clustering Variety (GFCV). The results from the fuzzy descriptor models are compared to the results of several other methods. An efficient technique, based on the error indices of multiple validation sets, is used to optimize the number of neurons and prevent the algorithm from over fitting. The scope of this work is to reveal the advantages of fuzzy descriptor models and compare them to the most successful neural and neuro fuzzy approaches based on prediction accuracy, generalization, and computational complexity. The proposed solution is shown to accurately forecast seismic time series, outperforming several other methods.